21:00:16 <timburke> #startmeeting swift 21:00:16 <opendevmeet> Meeting started Wed Aug 7 21:00:16 2024 UTC and is due to finish in 60 minutes. The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:16 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:16 <opendevmeet> The meeting name has been set to 'swift' 21:00:24 <timburke> who's here for the swift meeting? 21:00:35 <kota> hi 21:00:36 <mattoliver> o/ 21:00:49 <jianjian> o/ 21:01:28 <timburke> i neglected to update the agenda 21:01:49 <timburke> (seems like it takes me longer and longer to reload context after a vacation...can't decide whether that's because my vacations are getting better or i'm just getting older) 21:02:09 <mattoliver> Lol 21:02:12 <timburke> but there are definitely some updates i can give on the previous meeting's agenda 21:02:32 <timburke> #topic pkg_resources warnings and editable installs 21:02:52 <jianjian> ;-) 21:03:15 <timburke> all the patches to move to entry_points (instead of scripts) have landed! 21:03:38 <mattoliver> \o/ 21:03:50 <timburke> so nothing under bin/ gets used when building wheels or running `pip install -e .` 21:04:52 <timburke> the only remaining question is what to do with the bin scripts now. i kept them as i was moving code around, but i'm not sure they're actually *necessary* 21:05:15 <timburke> so i proposed 21:05:17 <timburke> #link https://review.opendev.org/c/openstack/swift/+/924795 21:05:17 <patch-bot> patch 924795 - swift - Remove legacy bin/ scripts - 3 patch sets 21:06:50 <timburke> i *think* the only reasonably valid use for them now is to not break developer work-flows -- but even there, i think we ought to be recommending that people use an editable install and get commands installed on your $PATH 21:07:24 <mattoliver> Didn't we decide a few meetings ago to wait a release or something? Or am I miss remembering. Personally I'm on the fence, it would be nice to get rid of them :) 21:08:54 <mattoliver> But even if we did, we can revisit. What do people think? Shall we just finish cleaning house? 21:09:01 <timburke> oh, maybe there was a decision already... i suppose i'm still partial toward ripping the band-aid off sooner rather than later ;-) 21:09:15 <mattoliver> Esp since their not used anymore 21:09:39 <mattoliver> Nothing wrong with us changing our minds. 21:09:58 <timburke> mainly i was going to suggest that people leave a comment on the review -- if there are a fair number of +1s and no -1s, let it merge 21:10:47 <mattoliver> Know that all the others have landed, I tested each patch and compared, and I kinda love how the mains now live where they're suppose to, so still easy to run from an ide 21:11:03 <mattoliver> Oh nice idea! 21:13:10 <timburke> i could also try to dig into some downstream packaging stuff and try to get a sense of whether they're used there (though i suspect they're not) 21:13:29 <timburke> for example, i'm pretty sure the debian stuff lives at https://salsa.debian.org/openstack-team/services/swift/-/tree/debian/caracal/debian?ref_type=heads 21:13:56 <timburke> i think i found centos spec files at some point, too, though i forget where now 21:14:38 <mattoliver> I'll find that one, I know where they live and comment in the patch if you like 21:14:52 <mattoliver> Just need to be by my computer 21:15:03 <timburke> but anyway, yeah -- please leave a +1/-1 on the review, and maybe we can talk it out a little more there 21:15:40 <timburke> #topic py313 bring-up 21:16:00 <timburke> a couple of the patches that came out of that merged already! 21:16:11 <timburke> #link https://review.opendev.org/c/openstack/swift/+/924768 21:16:11 <patch-bot> patch 924768 - swift - Implement context manager protocol for logging mut... (MERGED) - 2 patch sets 21:16:16 <timburke> #link https://review.opendev.org/c/openstack/swift/+/924771 21:16:16 <patch-bot> patch 924771 - swift - Pass timeout as kwarg (MERGED) - 2 patch sets 21:16:22 <timburke> thanks acoles! 21:16:56 <timburke> the patch to stop using the (now removed) cgi module has also seen some more back-and-forth 21:16:58 <mattoliver> Oh nice 21:17:14 <timburke> #link https://review.opendev.org/c/openstack/swift/+/887908 21:17:14 <patch-bot> patch 887908 - swift - Stop using cgi.parse_header - 4 patch sets 21:17:56 <timburke> (though maybe not since last meeting) but it led me to try out another approach in 21:17:59 <timburke> #link https://review.opendev.org/c/openstack/swift/+/925825 21:17:59 <patch-bot> patch 925825 - swift - Add helpers for adding/popping parameters - 1 patch set 21:19:00 <jianjian> good direction to add some helper functions 21:19:16 <mattoliver> I looked at that one yesterday 21:19:44 <mattoliver> Hopefully I hut submit on the review :😜 21:20:22 <timburke> as i dig into the test failures, though, i'm starting to smell some bugs on master... i think we might keep extending the container-update-etag when we copy symlinks, and there seems to be something fishy with static symlinks and object-versioning-enabled containers 21:20:26 <jianjian> I looked the previous patch, felt the reuse of parse_content_type was okay functionality wise, but that name is kind of annoying 😄 21:22:08 <jianjian> Tim, are those failures only limited to py27? 21:23:06 <timburke> nope -- there's legit func test failures across the board -- the validation i added in the append_header_parameter function is getting tripped (specifically, the duplicate parameter check) 21:23:47 <jianjian> okay, it's a real concern then 21:24:22 <timburke> now, as i'm digging into it, i could take the helpers patch and have it be a follow-up to the parse_content_type patch -- it'd give us the option of getting rid of our cgi usage faster, but grows the API of parse_content_type (which i'm starting to feel like we should maybe move away from)... 21:24:54 <jianjian> great! 21:25:21 <jianjian> need to drop from this meeting, I will take a look at this new patch later this week. 21:25:30 <timburke> k -- thanks jianjian! 21:26:44 <mattoliver> well if the helpers are in the pipeline, then maybe we can let parse_content_type do more for a short amount of time.. so long as we get the follow up landed at some point. 21:27:18 <mattoliver> I did suggest just renaming parse_content_type and keeping the old name around to for backwards compat. 21:27:47 <mattoliver> but having helper methods seems like a cleaner approach 21:29:16 <timburke> yeah... but i don't know that i actually *like* the parse_content_type API -- it seems inherently lossy and incomplete 21:31:46 <timburke> at least, given the way we always send the parameter list to a dict :-/ 21:32:42 <timburke> my worry about turning it into a follow-up is that there will never be enough impetus to actually drive it to completion 21:33:31 <mattoliver> yup 21:33:41 <mattoliver> because we fall into that trap alot. 21:33:46 <timburke> idk -- i'll at least push on it long enough to get the jobs passing 21:33:51 <mattoliver> so maybe we shold just do the right thing first 21:34:23 <timburke> all right, i think that's all i've got -- i still need to follow-up on that node-exclusion thread... 21:35:09 <timburke> #topic open discussion 21:35:17 <timburke> anything else we should talk about this week? 21:36:46 <mattoliver> I wrote a report on the qatar mentoring for the university they're from. Foundation are now thinking about anonymising it and asking about making it an article... so that's kinda surprising.. but will be about swift 21:37:05 <timburke> oh, cool! good job, mattoliver! 21:37:59 <mattoliver> Otherwise I have been working on stuff, but can't think of anything at the moment, so I got nothing.. oh I guess I should loop back to the qatar patches and give them some love 21:40:01 <mattoliver> thats all i got :) 21:40:24 <timburke> yeah, me too -- i saw acoles proposed something of a follow-up 21:40:28 <timburke> #link https://review.opendev.org/c/openstack/swift/+/925890 21:40:28 <patch-bot> patch 925890 - swift - tests: don't mutate and re-use Response headers - 2 patch sets 21:41:18 <mattoliver> oh cool, I missed that 21:41:25 <timburke> i should go approve that -- didn't see it until after i'd already proposed p 925900 after looking at the recent merge from master for feature/mpu 21:41:25 <patch-bot> https://review.opendev.org/c/openstack/swift/+/925900 - swift - tests: Explicitly register new account headers - 1 patch set 21:42:23 <timburke> all right, then i'll call it a little early and let you get on with your morning mattoliver :-) 21:42:28 <timburke> #endmeeting